Subphotospheric dissipation in gamma-ray bursts observed by the Fermi Gamma-ray Space Telescope

Gamma-ray bursts (GRBs) are the brightest events in the Universe, for a short time outshining the rest of the Universe combined, as they explode with isotropic equivalent luminosities up to $10^{54}$ erg s$^{-1}$. These events are believed to be connected to supernovae and to binary compact object mergers, such as binary neutron stars or neutron star -- black hole systems. The origin of the so-called prompt emission in GRBs remains an unsolved problem, although some progress is being made. Spectral analysis of prompt emission has traditionally been performed with the Band function, an empirical model with no physical interpretation, and it is just recently that physical models have started to be fitted to data. This thesis presents spectral analysis of GRB data from the Fermi Gamma-ray Space Telescope using a physical model for subphotospheric dissipation. The model is developed using a numerical code and implemented as a table model in {\scriptsize XSPEC}. Paper \rom{1} presents the model and provides a proof-of-concept of fitting GRB data with such a model. Specifically, two GRBs are fitted and compared with the corresponding Band function fits. In paper \rom{2}, a sample of 37 bursts are fitted with an extended version of the model and improved analysis tools. Overall, about a third of the fitted spectra can be described by the model. From these fits it is concluded that the scenario of subphotospheric dissipation can describe all spectral shapes present in the sample. The key characteristic of the spectra that are not fitted by the model is that they are very luminous. Within the context of the model, this suggests that the assumption of internal shocks as a dissipation mechanism cannot explain the full population of GRBs. Alternatively, additional emission components may required. The thesis concludes that subphotospheric dissipation is viable as a possible origin of GRB prompt emission. Furthermore, it shows the importance of using physically motivated models when analysing GRBs.QC 20170221